Jump to content

Interpreting SMART readings and preclear results

Recommended Posts

I know there is a thread dedicated for preclear results. I have posted in there but it does not appear to get much attention. I gather that the main concern with preclear is that you get the section that reports on any reallocated sectors as being zero all the way through if possible. Mine do state that. However as I posted in the other thread and below, I am not sure if SMART output data should be read as simply as "if nothing states a failure then nothing is wrong".



"Update and some questions. I have finished preclears for all three of my drives. The part where it mentions in the results about "no Smart attribultes are failing now", has NEVER failed. i.e. All of the drives during all three preclears each have always had just zeros with no reallocated sectors. So that's something.


However, the smart testing sections differ somewhat. As I mentioned above, sometimes they give raw_read_error_rate numbers, and sometimes they remain at zero, and also I got the Samsung drive on the last preclear complain of a multi_zone_error_rate = 1.


How are we meant to decipher all of these Smart read outs easily other than posting here and asking? I am becoming a little bit concerned that perhaps unraid involves a lot of "being paranoid" over drive reports. A quick google about any of the things smart reads and it throughs up a load of paranoia about a possibility of the drive failing and general recommendation to RMA it. I see that my drives have a varying load_cycle_count of bewteen 8 for the Hitachi drives, and 120 for the Samsung. These drives have only been on about 3 - 4 days. Should I be concerned of a load cycle count of 120 83 hours? That is 1.4 times an hour on average. That's about 12000 load cycles per year. Which would equate to it potentially failing in under 3 years if one is to assume the load cycle rating is approximately 30000 which seems to be a recommended expectation based on a quick google of other similar spec 2TB drives.


So I could use some advice on the multi zone error rate, load cycle and raw_read_error rate stuff and whether to proceed with my build basically. Or even if I am just reading it all wrong and or over the top.




Link to comment

As you said, there's a lot of paranoia surrounding SMART reports.  I choose to trust unRAID to protect my data, and I generally trust HDD manufacturers to produce quality products (though I do my research before buying a new drive).  The only time I'll ever fret is if I see the signs of two or more drives failing at once - a condition from which unRAID cannot recover.  If you want to be extra-cautious, keep a spare precleared drive handy so that you can replace a single failed drive at any time with minimal downtime.


As for preclear reports, as long as you see all zeros in the 'differences' section, then it is generally fine to just trust the drive and move on with life.  If you want to learn more about each SMART parameter feel free to do so, but keep in mind that the majority of them are meaningless to anyone but the drive manufacturer.  I think Joe L. did a great job in simplifying the SMART output into the 'differences' report.  That report notes the most significant SMART attributes that a consumer can meaningfully interpret.


Many drives will throw one or two errors in a particular SMART parameter but for all intents and purposes still be a perfectly healthy and trustworthy drive.  If you see one or two errors in an initial SMART report then take note of them, but don't freak out.  If you see a particular SMART parameter constantly increasing at an alarming rate, then of course that is something to closely monitor.


By the way, it is perfectly normal for load cycle counts to vary between drive models and manufacturers, and especially between 7200rpm and 'green' drives.  One of the methods that manufacturers use to make a drive more energy efficient is to park the head more often, which will result in increased load cycle counts.  As long as these values aren't skyrocketing towards the threshold, then you have nothing to worry about.


I hope this helps put your mind at ease.

Link to comment

It does put my mind at ease, and I apologise if I came accross as "frustrated" reading my post back. I will crack on with the build although I have more questions! Namely about NFS vs SMB, user share/data layout, copying data to array etc. I will read up.

Link to comment


This topic is now archived and is closed to further replies.

  • Create New...